# Modules & Libraries
import pandas as pd
import numpy as np
import matplotlib
import matplotlib.pyplot as plt
import seaborn as sns
from matplotlib.dates import DateFormatter, MonthLocator
import matplotlib.dates as mdates
import copy
import umap.umap_ as umap
from sklearn.preprocessing import StandardScaler
from matplotlib.animation import FuncAnimation

# Loading the dataset
path = '../Data/hotel_bookings.csv'

# Read in csv file
df = pd.read_csv(path)
df.head()

# Note that the describe function defaults to using only some variables
df.describe()

# Returns the dimensions of the DataFrame as (number of rows, number of columns)
df.shape

(119390, 32)

# Returns an index object containing the column labels of the DataFrame.
df.columns

Index(['hotel', 'is_canceled', 'lead_time', 'arrival_date_year',
       'arrival_date_month', 'arrival_date_week_number',
       'arrival_date_day_of_month', 'stays_in_weekend_nights',
       'stays_in_week_nights', 'adults', 'children', 'babies', 'meal',
       'country', 'market_segment', 'distribution_channel',
       'is_repeated_guest', 'previous_cancellations',
       'previous_bookings_not_canceled', 'reserved_room_type',
       'assigned_room_type', 'booking_changes', 'deposit_type', 'agent',
       'company', 'days_in_waiting_list', 'customer_type', 'adr',
       'required_car_parking_spaces', 'total_of_special_requests',
       'reservation_status', 'reservation_status_date'],
      dtype='object')

# Provides a concise summary of the DataFrame including data types, non-null values, and memory usage.
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 119390 entries, 0 to 119389
Data columns (total 32 columns):
 #   Column                          Non-Null Count   Dtype  
---  ------                          --------------   -----  
 0   hotel                           119390 non-null  object 
 1   is_canceled                     119390 non-null  int64  
 2   lead_time                       119390 non-null  int64  
 3   arrival_date_year               119390 non-null  int64  
 4   arrival_date_month              119390 non-null  object 
 5   arrival_date_week_number        119390 non-null  int64  
 6   arrival_date_day_of_month       119390 non-null  int64  
 7   stays_in_weekend_nights         119390 non-null  int64  
 8   stays_in_week_nights            119390 non-null  int64  
 9   adults                          119390 non-null  int64  
 10  children                        119386 non-null  float64
 11  babies                          119390 non-null  int64  
 12  meal                            119390 non-null  object 
 13  country                         118902 non-null  object 
 14  market_segment                  119390 non-null  object 
 15  distribution_channel            119390 non-null  object 
 16  is_repeated_guest               119390 non-null  int64  
 17  previous_cancellations          119390 non-null  int64  
 18  previous_bookings_not_canceled  119390 non-null  int64  
 19  reserved_room_type              119390 non-null  object 
 20  assigned_room_type              119390 non-null  object 
 21  booking_changes                 119390 non-null  int64  
 22  deposit_type                    119390 non-null  object 
 23  agent                           103050 non-null  float64
 24  company                         6797 non-null    float64
 25  days_in_waiting_list            119390 non-null  int64  
 26  customer_type                   119390 non-null  object 
 27  adr                             119390 non-null  float64
 28  required_car_parking_spaces     119390 non-null  int64  
 29  total_of_special_requests       119390 non-null  int64  
 30  reservation_status              119390 non-null  object 
 31  reservation_status_date         119390 non-null  object 
dtypes: float64(4), int64(16), object(12)
memory usage: 29.1+ MB

# Attributes Dropped
df.drop(['arrival_date_year', 'arrival_date_month', 
         'arrival_date_week_number', 
         'arrival_date_day_of_month', 
         'agent', 
         'company', 
         'required_car_parking_spaces',
         'meal',
         'country',
         'market_segment',
         'distribution_channel',
         'previous_bookings_not_canceled',
         'reserved_room_type',
         'assigned_room_type',
         'days_in_waiting_list',
         'required_car_parking_spaces',
         'reservation_status',
         'total_of_special_requests'
         ],
        axis=1,
        inplace=True)

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 119390 entries, 0 to 119389
Data columns (total 15 columns):
 #   Column                   Non-Null Count   Dtype  
---  ------                   --------------   -----  
 0   hotel                    119390 non-null  object 
 1   is_canceled              119390 non-null  int64  
 2   lead_time                119390 non-null  int64  
 3   stays_in_weekend_nights  119390 non-null  int64  
 4   stays_in_week_nights     119390 non-null  int64  
 5   adults                   119390 non-null  int64  
 6   children                 119386 non-null  float64
 7   babies                   119390 non-null  int64  
 8   is_repeated_guest        119390 non-null  int64  
 9   previous_cancellations   119390 non-null  int64  
 10  booking_changes          119390 non-null  int64  
 11  deposit_type             119390 non-null  object 
 12  customer_type            119390 non-null  object 
 13  adr                      119390 non-null  float64
 14  reservation_status_date  119390 non-null  object 
dtypes: float64(2), int64(9), object(4)
memory usage: 13.7+ MB

# Check for missing values
df.isnull().sum()

hotel                      0
is_canceled                0
lead_time                  0
stays_in_weekend_nights    0
stays_in_week_nights       0
adults                     0
children                   4
babies                     0
is_repeated_guest          0
previous_cancellations     0
booking_changes            0
deposit_type               0
customer_type              0
adr                        0
reservation_status_date    0
dtype: int64

# Counting the missing values per col/attribute
missing_counts = df.isnull().sum()

# Set figure size
plt.figure(figsize=(10, 6))

# Plot missing values
missing_counts[missing_counts > 0].plot(kind='bar')

# Set title labels, rotate x-axis
plt.title('Missing Values Count by Column')
plt.xlabel('Column')
plt.ylabel('Number of Missing Values')
plt.xticks(rotation=45)

# Display
plt.show()

# Checking for Duplicates
df.duplicated().sum()

np.int64(36256)

# Example: Last two entries here are duplicates!
df.head(6)

# Convert hotel to boolean
df['hotel'] = df['hotel'].map({'Resort Hotel': 0, 'City Hotel': 1}).astype('bool')

# Convert is_repeated_guest to boolean
df['is_repeated_guest'] = df['is_repeated_guest'].astype('bool')

# One hot encoding deposit_type, indicating presence (1) or absence (0).
if 'deposit_type' in df.columns:
    df = pd.get_dummies(df, columns=['deposit_type'], prefix=['deposit_type'])

# Imputing misisng values, children attribute with 0
df['children'] = df['children'].fillna(0).astype('int') 

# Converting reservation_status_date to month_year
df['reservation_status_date'] = pd.to_datetime(df['reservation_status_date'], format='%d-%m-%y', errors='coerce')

# Create month_year column in monthly period format
df['month_year'] = df['reservation_status_date'].dt.to_period('M')

# Filter Dataframe
start_period = pd.Period('2015-07', freq='M')
df = df[df['month_year'] >= start_period]

# Aggregation of 'adults', 'children', and 'babies' into 'num_of_guests'
df['num_of_guests'] = df['adults'] + df['children'] + df['babies']

# Aggregation of 'stays_in_weekend_nights' and 'stays_in_week_nights' into 'length_of_stay'
df['length_of_stay'] = df['stays_in_weekend_nights'] + df['stays_in_week_nights']

# Dropping the attributes we aggregated
df = df.drop(columns=['adults', 'children', 'babies', 'stays_in_weekend_nights', 'stays_in_week_nights'])

df.info()

<class 'pandas.core.frame.DataFrame'>
Index: 117040 entries, 0 to 119389
Data columns (total 15 columns):
 #   Column                   Non-Null Count   Dtype         
---  ------                   --------------   -----         
 0   hotel                    117040 non-null  bool          
 1   is_canceled              117040 non-null  int64         
 2   lead_time                117040 non-null  int64         
 3   is_repeated_guest        117040 non-null  bool          
 4   previous_cancellations   117040 non-null  int64         
 5   booking_changes          117040 non-null  int64         
 6   customer_type            117040 non-null  object        
 7   adr                      117040 non-null  float64       
 8   reservation_status_date  117040 non-null  datetime64[ns]
 9   deposit_type_No Deposit  117040 non-null  bool          
 10  deposit_type_Non Refund  117040 non-null  bool          
 11  deposit_type_Refundable  117040 non-null  bool          
 12  month_year               117040 non-null  period[M]     
 13  num_of_guests            117040 non-null  int64         
 14  length_of_stay           117040 non-null  int64         
dtypes: bool(5), datetime64[ns](1), float64(1), int64(6), object(1), period[M](1)
memory usage: 10.4+ MB

# No more null values!
df.isnull().sum()

hotel                      0
is_canceled                0
lead_time                  0
is_repeated_guest          0
previous_cancellations     0
booking_changes            0
customer_type              0
adr                        0
reservation_status_date    0
deposit_type_No Deposit    0
deposit_type_Non Refund    0
deposit_type_Refundable    0
month_year                 0
num_of_guests              0
length_of_stay             0
dtype: int64

# Dropping Duplicates
df.drop_duplicates(inplace=True)

# No more duplicates!
df.duplicated().sum()

np.int64(0)

df.info()

<class 'pandas.core.frame.DataFrame'>
Index: 82363 entries, 0 to 119389
Data columns (total 15 columns):
 #   Column                   Non-Null Count  Dtype         
---  ------                   --------------  -----         
 0   hotel                    82363 non-null  bool          
 1   is_canceled              82363 non-null  int64         
 2   lead_time                82363 non-null  int64         
 3   is_repeated_guest        82363 non-null  bool          
 4   previous_cancellations   82363 non-null  int64         
 5   booking_changes          82363 non-null  int64         
 6   customer_type            82363 non-null  object        
 7   adr                      82363 non-null  float64       
 8   reservation_status_date  82363 non-null  datetime64[ns]
 9   deposit_type_No Deposit  82363 non-null  bool          
 10  deposit_type_Non Refund  82363 non-null  bool          
 11  deposit_type_Refundable  82363 non-null  bool          
 12  month_year               82363 non-null  period[M]     
 13  num_of_guests            82363 non-null  int64         
 14  length_of_stay           82363 non-null  int64         
dtypes: bool(5), datetime64[ns](1), float64(1), int64(6), object(1), period[M](1)
memory usage: 7.3+ MB

# Reset the index just so it's easier to read the rows/entries
df = df.reset_index(drop=True)
df.tail()

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 82363 entries, 0 to 82362
Data columns (total 15 columns):
 #   Column                   Non-Null Count  Dtype         
---  ------                   --------------  -----         
 0   hotel                    82363 non-null  bool          
 1   is_canceled              82363 non-null  int64         
 2   lead_time                82363 non-null  int64         
 3   is_repeated_guest        82363 non-null  bool          
 4   previous_cancellations   82363 non-null  int64         
 5   booking_changes          82363 non-null  int64         
 6   customer_type            82363 non-null  object        
 7   adr                      82363 non-null  float64       
 8   reservation_status_date  82363 non-null  datetime64[ns]
 9   deposit_type_No Deposit  82363 non-null  bool          
 10  deposit_type_Non Refund  82363 non-null  bool          
 11  deposit_type_Refundable  82363 non-null  bool          
 12  month_year               82363 non-null  period[M]     
 13  num_of_guests            82363 non-null  int64         
 14  length_of_stay           82363 non-null  int64         
dtypes: bool(5), datetime64[ns](1), float64(1), int64(6), object(1), period[M](1)
memory usage: 6.7+ MB

df.describe()

# Resolving the negative value in adr
df = df[df['adr'] >= 0]

# IQR Function
def remove_outliers_iqr(df, column):
    # Calculate the first and third quartile
    Q1 = df[column].quantile(0.25)
    Q3 = df[column].quantile(0.75)

    ## Compute the Interquartile Range (IQR)
    IQR = Q3 - Q1

    # Define lower and upper bounds for outlier detection
    lower_bound = Q1 - 1.5 * IQR
    upper_bound = Q3 + 1.5 * IQR

    # Filter the DataFrame to remove outliers
    df= df[(df[column] >= lower_bound) & (df[column] <= upper_bound)]
    return df

# Define columns for which outlier to remove
columns_to_filter = ['lead_time', 'adr']

# Iterate and apply function
for col in columns_to_filter:
    df = remove_outliers_iqr(df, col)

df = df.reset_index(drop=True)

# Extreme outliers from `lead_time` and `adr` have been removed! 
df.describe()

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 78017 entries, 0 to 78016
Data columns (total 15 columns):
 #   Column                   Non-Null Count  Dtype         
---  ------                   --------------  -----         
 0   hotel                    78017 non-null  bool          
 1   is_canceled              78017 non-null  int64         
 2   lead_time                78017 non-null  int64         
 3   is_repeated_guest        78017 non-null  bool          
 4   previous_cancellations   78017 non-null  int64         
 5   booking_changes          78017 non-null  int64         
 6   customer_type            78017 non-null  object        
 7   adr                      78017 non-null  float64       
 8   reservation_status_date  78017 non-null  datetime64[ns]
 9   deposit_type_No Deposit  78017 non-null  bool          
 10  deposit_type_Non Refund  78017 non-null  bool          
 11  deposit_type_Refundable  78017 non-null  bool          
 12  month_year               78017 non-null  period[M]     
 13  num_of_guests            78017 non-null  int64         
 14  length_of_stay           78017 non-null  int64         
dtypes: bool(5), datetime64[ns](1), float64(1), int64(6), object(1), period[M](1)
memory usage: 6.3+ MB

# Cross Tabulation Showing Cancellation Counts Among Numerical Attributes
# List of numerical columns to analyze
numerical = ['lead_time', 'previous_cancellations', 'booking_changes', 'adr', 'num_of_guests', 'length_of_stay']

# Create copy of df to not modify original data
crosstab_df = copy.deepcopy(df)
crosstab_df['is_canceled'] = df['is_canceled'].map({0: 'Not Canceled', 1: 'Canceled'})

# Num of numerical columns
L = len(numerical)

# Num of coloumns in grid and calculate num of rows
ncol = 3
nrow = int(np.ceil(L / ncol))

# Setup matplotlib and axis
fig, axes = plt.subplots(nrow, ncol, figsize=(15, 5 * nrow))
fig.subplots_adjust(top=0.92)
axes = axes.flatten()

# Loop over numerical column to create histogram and stack
for i, col in enumerate(numerical):
    sns.histplot(data=crosstab_df, x=col, hue="is_canceled", multiple="stack", ax=axes[i], bins=30, palette="coolwarm")
    # Set title, labels, enable grid
    axes[i].set_title(f'Distribution of {col}', fontsize=14)
    axes[i].set_xlabel(f'{col} Value', fontsize=12)
    axes[i].set_ylabel('Frequency', fontsize=12)
    axes[i].grid(True, which='both', linestyle='--', linewidth=0.5)
    sns.despine(ax=axes[i])

# If empty subplots delete axes
for j in range(L, nrow * ncol):
    fig.delaxes(axes[j])

# Cleaner look and display
plt.tight_layout()
plt.show()

# Scatterplot of Lead Time vs ADR with Cancellation Status

# Create copy of df to not modify original data and map binary cancellation w/ readable labels
scatter_copy = copy.deepcopy(df)
scatter_copy['is_canceled'] = scatter_copy['is_canceled'].map({0: 'Not Canceled', 1: 'Canceled'})

# Set figure size
plt.figure(figsize=(12, 7))

# Create scatterplot lead time vs ADR
sns.scatterplot(data=scatter_copy, x='lead_time', y='adr', hue='is_canceled', s=50, alpha=0.7, palette='pastel')

# Set title, labels, add legend
plt.title('Scatterplot of Lead Time vs ADR with Cancellation Status')
plt.xlabel('Lead Time')
plt.ylabel('Average Daily Rate (ADR)')
plt.legend(title='Cancellation Status')

# Cleaner look and display
plt.tight_layout()
plt.show()

# Box Plot Displaying the Correlation between Length of Stay and Cancellation Status
# Set figure size
plt.figure(figsize=(10, 6))

# Create box plot to display correlation length of stay and cancellation
ax = sns.boxplot(x='is_canceled', y='length_of_stay', data=df, 
                 hue='is_canceled', palette='pastel')

# Create dictionary to map cancellation status
labels = {0: 'Not Canceled', 1: 'Canceled'}

# Set the xaxis, and assign labels
ax.set_xticks([0, 1]) 
ax.set_xticklabels([labels[0], labels[1]])

# Set titles and labels, remove the legend
plt.xlabel('Cancellation Status')
plt.ylabel('Length of Stay')
plt.title('Length of Stay by Cancellation Status')
ax.legend_.remove()

# Display
plt.show()

# Distribution of Average Daily Rate
# Set style of seaborn plot and context
sns.set_style("darkgrid")
sns.set_context("talk")  

# Set figure size
plt.figure(figsize=(10, 6))

# Create Histogram with KDE
sns.histplot(data=df, x='adr', kde=True, bins=20, color='skyblue', line_kws={'color': 'darkblue', 'lw': 3})

# Set title, labels, grid lines
plt.title('Distribution of Average Daily Rate', fontsize=16, weight='bold')
plt.xlabel('Average Daily Rate', fontsize=14)
plt.ylabel('Frequency', fontsize=14)
plt.grid(True, linestyle='--', alpha=0.7)

# Display
plt.show()

# Statistics
df['adr'].describe()

count    78017.000000
mean       103.314852
std         46.218942
min          0.000000
25%         72.250000
50%         98.000000
75%        132.000000
max        230.000000
Name: adr, dtype: float64

# Correlation Heatmap of Attributes
# Create copy of df to not modify original data
df_temp = copy.deepcopy(df)

# Converting boolean columns to integers in the temporary DataFrame
bool_cols = ['hotel', 'is_repeated_guest', 'deposit_type_No Deposit', 'deposit_type_Non Refund', 'deposit_type_Refundable']
df_temp[bool_cols] = df_temp[bool_cols].astype(int)

# Select column for correlation analysis, and compute
correlation_cols = df_temp[['is_canceled', 'lead_time', 'previous_cancellations', 'booking_changes', 'adr', 'num_of_guests', 'length_of_stay'] + bool_cols]
correlation_matrix = correlation_cols.corr()

# Set figure size
plt.figure(figsize=(12, 10))

# Create heatmap
sns.heatmap(correlation_matrix, annot=True, cmap="viridis", linewidths=0.5, fmt=".2f")

# Set title and display
plt.title('Correlation Heatmap of Attributes')
plt.show()

# Pie Chart Displaying Cancellation Rates for Each Deposit Type
deposit_grouped = df.groupby(
    ['deposit_type_No Deposit', 'deposit_type_Non Refund', 'deposit_type_Refundable', 'is_canceled']
).size().unstack(fill_value=0)

# Aggregating cancellations  for each deposit type
cancellation_counts = deposit_grouped.loc[:, 1]

# Calculating cancellations for each deposit type directly
cancellation_no_deposit = df[(df['deposit_type_No Deposit'] == True) & (df['is_canceled'] == 1)].shape[0]
cancellation_non_refund = df[(df['deposit_type_Non Refund'] == True) & (df['is_canceled'] == 1)].shape[0]
cancellation_refundable = df[(df['deposit_type_Refundable'] == True) & (df['is_canceled'] == 1)].shape[0]

# Updating cancellation_counts with recalculated values
cancellation_counts = pd.Series({
    'No Deposit': cancellation_no_deposit,
    'Non Refund': cancellation_non_refund,
    'Refundable': cancellation_refundable
})

# Calculating percentages
total_cancellations = cancellation_counts.sum()
percentages = (cancellation_counts / total_cancellations) * 100

# Creating labels for the legend with percentages
legend_labels = [f'{label} - {percent:.1f}%' for label, percent in zip(cancellation_counts.index, percentages)]
colors = ['#ff9999', '#66b3ff', '#99ff99']

# Creating the pie chart without labels and percentages on the pie itself
plt.figure(figsize=(10, 7))
patches, _ = plt.pie(cancellation_counts, startangle=90, colors=colors, wedgeprops={'edgecolor': 'black'}, labels=None)

# Adjusting legend to include percentages
plt.legend(patches, legend_labels, loc='center left', bbox_to_anchor=(1, 0.5), fontsize=8, title='Deposit Type')
plt.title('Cancellations by Deposit Type')
plt.show()

# Displaying the number of cancellations for each deposit type
print('Number of Cancellations by Deposit Type:')
cancellation_counts

Number of Cancellations by Deposit Type:

No Deposit    20554
Non Refund      654
Refundable       23
dtype: int64

# Cancellation Rates by Repeat Status in Each Hotel Type
# Group data by repeated guest status/hotel type and calculate mean
cancellation_rate = df.groupby(['is_repeated_guest', 'hotel'])['is_canceled'].mean().reset_index()

# Mapping boolean values to appropriate labels
cancellation_rate['is_repeated_guest'] = cancellation_rate['is_repeated_guest'].map({True: 'Repeated', False: 'Not Repeated'})
cancellation_rate['hotel'] = cancellation_rate['hotel'].map({True: 'City Hotel', False: 'Resort Hotel'})

# Setting plot style and context with custom colors
sns.set_style("darkgrid")
sns.set_context("talk")
colors = ['#66c2a5', '#fc8d62']  

# Create a bar plot to visualize cancellation rate
plt.figure(figsize=(10, 6))
sns.barplot(x='is_repeated_guest', y='is_canceled', hue='hotel', data=cancellation_rate, palette=colors)

# Set title and labels
plt.title('Cancellation Rate by Repeated Guest Status and Hotel Type', fontsize=16, weight='bold')
plt.xlabel('Repeated Guest Status', fontsize=14)
plt.ylabel('Cancellation Rate', fontsize=14)

# Customize legend, add grid for readability
plt.legend(title='Hotel Type', fontsize=12, title_fontsize='13')
plt.grid(True, linestyle='--', alpha=0.7)
plt.show()

# Output cancellation rate
cancellation_rate

# Relationship of month_year and cancellations across the two types of hotels
monthly_cancellations = df.groupby(['month_year', 'hotel'])['is_canceled'].mean().reset_index()

# Converting 'month_year' to string format for plotting
monthly_cancellations['month_year'] = monthly_cancellations['month_year'].astype(str)
monthly_cancellations['hotel'] = monthly_cancellations['hotel'].map({True: 'City Hotel', False: 'Resort Hotel'})

# Create line plot
plt.figure(figsize=(12, 6))
sns.lineplot(data=monthly_cancellations, x='month_year', y='is_canceled', hue='hotel', marker='o')

# Rotate x-axis, add labels, and title
plt.xticks(rotation=45)
plt.xlabel('Month-Year')
plt.ylabel('Average Cancellation Rate')
plt.title('Monthly Average Cancellation Rate by Hotel Type')

# Customize legend
plt.legend(title='Hotel Type')

# Clean and display
plt.tight_layout()
plt.show()

def cancellation_by_customer_type(df):
    """
    Function to calculate the cancellation rate by customer type.
    
    Args:
    df (pandas.DataFrame): DataFrame containing the columns 'customer_type' and 'is_canceled'.
    
    Returns:
    pandas.DataFrame: A DataFrame with customer types and their corresponding cancellation rates.
    """
    
    # Group by customer_type and calculate the cancellation rate
    cancellation_rate = df.groupby('customer_type')['is_canceled'].mean().reset_index()
    
    # Rename the columns for better readability
    cancellation_rate.columns = ['customer_type', 'cancellation_rate']
    
    # Convert cancellation rate to percentage for easier interpretation
    cancellation_rate['cancellation_rate'] = cancellation_rate['cancellation_rate'] * 100
    
    return cancellation_rate


result = cancellation_by_customer_type(df)
print(result)

     customer_type  cancellation_rate
0         Contract          15.224417
1            Group           6.490872
2        Transient          29.133241
3  Transient-Party          17.274130

def plot_violin_cancellation_by_customer_type(df):
    """
    Function to plot a violin graph showing the distribution of cancellations by customer type.
    
    Args:
    df (pandas.DataFrame): DataFrame containing the columns 'customer_type' and 'is_canceled'.
    """
    # Set up the plot size and style
    plt.figure(figsize=(10, 6))
    sns.set(style="whitegrid")

    # Create the violin plot
    sns.violinplot(x='customer_type', y='is_canceled', data=df, inner="point", hue='customer_type', palette="muted", legend=False)
    
    # Add title and labels
    plt.title('Distribution of Cancellations by Customer Type', fontsize=16)
    plt.xlabel('Customer Type', fontsize=12)
    plt.ylabel('Cancellation (0 = No, 1 = Yes)', fontsize=12)
    
    # Show the plot
    plt.show()

plot_violin_cancellation_by_customer_type(df)

# UMAP
# Integer columns for UMAP
int_columns = ['is_canceled', 'lead_time', 'previous_cancellations', 'booking_changes', 'num_of_guests', 'length_of_stay']
data = df[int_columns]

# Standardizing the data
scaler = StandardScaler()
data_scaled = scaler.fit_transform(data)

# Initializing UMAP model
reducer = umap.UMAP(n_neighbors=15, min_dist=0.1, n_components=2)

# Applying UMAP
embedding = reducer.fit_transform(data_scaled)

# Set figure size, 
plt.figure(figsize=(10, 7))

# Creat scatterplot of UMAP, add colorbar
plt.scatter(embedding[:, 0], embedding[:, 1], c=df['is_canceled'], cmap='Spectral', s=5, alpha=0.5)
plt.colorbar(label='Is Canceled')

# Set titles, labels
plt.title('UMAP Projection of Hotel Reservation Data')
plt.xlabel('UMAP Dimension 1')
plt.ylabel('UMAP Dimension 2')

# Display
plt.show()

# UMAP
int_columns = ['is_canceled', 'lead_time', 'previous_cancellations', 'booking_changes', 'num_of_guests', 'length_of_stay']
data = df[int_columns]

# Standardizing the data
scaler = StandardScaler()
data_scaled = scaler.fit_transform(data)

# Initialize UMAP model
reducer = umap.UMAP(n_neighbors=15, min_dist=0.1, n_components=2)

# Precomputing embeddings for each frame by slightly varying the min_dist parameter
embeddings = []
for frame in range(20):  # Reducing the number of frames to 20
    reducer = umap.UMAP(n_neighbors=15, min_dist=0.1 + 0.01 * frame, n_components=2)
    embedding = reducer.fit_transform(data_scaled)
    embeddings.append(embedding)

# Setting up the figure for animation
fig, ax = plt.subplots(figsize=(10, 7))

# Initial scatter plot
sc = ax.scatter(embeddings[0][:, 0], embeddings[0][:, 1], c=df['is_canceled'], cmap='Spectral', s=5, alpha=0.5)
plt.colorbar(sc, label='Is Canceled')
plt.title('UMAP Projection of Hotel Reservation Data')
plt.xlabel('UMAP Dimension 1')
plt.ylabel('UMAP Dimension 2')

# Function to update the plot for each frame
def update(frame):
    ax.clear()
    
    # Update scatter plot with precomputed embeddings
    embedding = embeddings[frame]
    sc = ax.scatter(embedding[:, 0], embedding[:, 1], c=df['is_canceled'], cmap='Spectral', s=5, alpha=0.5)
    plt.title(f'UMAP Projection of Hotel Reservation Data (Frame {frame})')
    plt.xlabel('UMAP Dimension 1')
    plt.ylabel('UMAP Dimension 2')
    
    return sc,

# Creating the animation
ani = FuncAnimation(fig, update, frames=20, interval=300)

# Displaying the animation in the notebook
from IPython.display import HTML
HTML(ani.to_jshtml())
# Saving as GIF 
# ani.save('umap_animation.gif', writer='imagemagick')

c:\Users\Juan Dominguez\AppData\Local\Programs\Python\Python312\Lib\site-packages\sklearn\manifold\_spectral_embedding.py:329: UserWarning: Graph is not fully connected, spectral embedding may not work as expected.
  warnings.warn(
c:\Users\Juan Dominguez\AppData\Local\Programs\Python\Python312\Lib\site-packages\umap\spectral.py:550: UserWarning: Spectral initialisation failed! The eigenvector solver
failed. This is likely due to too small an eigengap. Consider
adding some noise or jitter to your data.

Falling back to random initialisation!
  warn(
c:\Users\Juan Dominguez\AppData\Local\Programs\Python\Python312\Lib\site-packages\umap\spectral.py:550: UserWarning: Spectral initialisation failed! The eigenvector solver
failed. This is likely due to too small an eigengap. Consider
adding some noise or jitter to your data.

Falling back to random initialisation!
  warn(

	is_canceled	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	previous_cancellations	previous_bookings_not_canceled	booking_changes	agent	company	days_in_waiting_list	adr	required_car_parking_spaces	total_of_special_requests
count	119390.000000	119390.000000	119390.000000	119390.000000	119390.000000	119390.000000	119390.000000	119390.000000	119386.000000	119390.000000	119390.000000	119390.000000	119390.000000	119390.000000	103050.000000	6797.000000	119390.000000	119390.000000	119390.000000	119390.000000
mean	0.370416	104.011416	2016.156554	27.165173	15.798241	0.927599	2.500302	1.856403	0.103890	0.007949	0.031912	0.087118	0.137097	0.221124	86.693382	189.266735	2.321149	101.831122	0.062518	0.571363
std	0.482918	106.863097	0.707476	13.605138	8.780829	0.998613	1.908286	0.579261	0.398561	0.097436	0.175767	0.844336	1.497437	0.652306	110.774548	131.655015	17.594721	50.535790	0.245291	0.792798
min	0.000000	0.000000	2015.000000	1.000000	1.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	1.000000	6.000000	0.000000	-6.380000	0.000000	0.000000
25%	0.000000	18.000000	2016.000000	16.000000	8.000000	0.000000	1.000000	2.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	9.000000	62.000000	0.000000	69.290000	0.000000	0.000000
50%	0.000000	69.000000	2016.000000	28.000000	16.000000	1.000000	2.000000	2.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	14.000000	179.000000	0.000000	94.575000	0.000000	0.000000
75%	1.000000	160.000000	2017.000000	38.000000	23.000000	2.000000	3.000000	2.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	229.000000	270.000000	0.000000	126.000000	0.000000	1.000000
max	1.000000	737.000000	2017.000000	53.000000	31.000000	19.000000	50.000000	55.000000	10.000000	10.000000	1.000000	26.000000	72.000000	21.000000	535.000000	543.000000	391.000000	5400.000000	8.000000	5.000000

	is_canceled	lead_time	previous_cancellations	booking_changes	adr	reservation_status_date	num_of_guests	length_of_stay
count	82363.000000	82363.000000	82363.000000	82363.000000	82363.000000	82363	82363.000000	82363.000000
mean	0.280672	78.356434	0.026468	0.273120	107.326994	2016-09-07 04:31:01.826305536	2.039265	3.661280
min	0.000000	0.000000	0.000000	0.000000	-6.380000	2015-07-01 00:00:00	0.000000	0.000000
25%	0.000000	11.000000	0.000000	0.000000	72.750000	2016-03-25 00:00:00	2.000000	2.000000
50%	0.000000	48.000000	0.000000	0.000000	99.000000	2016-09-13 00:00:00	2.000000	3.000000
75%	1.000000	123.000000	0.000000	0.000000	135.000000	2017-03-08 00:00:00	2.000000	5.000000
max	1.000000	737.000000	13.000000	21.000000	5400.000000	2017-09-14 00:00:00	20.000000	69.000000
std	0.449330	84.411887	0.298818	0.736445	55.641959	NaN	0.712955	2.785946

	hotel	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_week_nights	adults	...	deposit_type	agent	company	customer_type	adr	total_of_special_requests	reservation_status	reservation_status_date
0	Resort Hotel	342	2015	July	27	1	0	2	...	No Deposit	NaN	NaN	Transient	0.0	0	Check-Out	01-07-15
1	Resort Hotel	737	2015	July	27	1	0	2	...	No Deposit	NaN	NaN	Transient	0.0	0	Check-Out	01-07-15
2	Resort Hotel	7	2015	July	27	1	1	1	...	No Deposit	NaN	NaN	Transient	75.0	0	Check-Out	02-07-15
3	Resort Hotel	13	2015	July	27	1	1	1	...	No Deposit	304.0	NaN	Transient	75.0	0	Check-Out	02-07-15
4	Resort Hotel	14	2015	July	27	1	2	2	...	No Deposit	240.0	NaN	Transient	98.0	1	Check-Out	03-07-15

	hotel	lead_time	is_repeated_guest	customer_type	adr	reservation_status_date	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	month_year	num_of_guests	length_of_stay
82358	True	23	False	Transient	96.14	2017-09-06	True	False	False	2017-09	2	7
82359	True	102	False	Transient	225.43	2017-09-07	True	False	False	2017-09	3	7
82360	True	34	False	Transient	157.71	2017-09-07	True	False	False	2017-09	2	7
82361	True	109	False	Transient	104.40	2017-09-07	True	False	False	2017-09	2	7
82362	True	205	False	Transient	151.20	2017-09-07	True	False	False	2017-09	2	9

	is_canceled	lead_time	previous_cancellations	booking_changes	adr	reservation_status_date	num_of_guests	length_of_stay
count	78017.000000	78017.000000	78017.000000	78017.000000	78017.000000	78017	78017.000000	78017.000000
mean	0.272133	71.019803	0.026751	0.263763	103.314852	2016-09-01 20:40:41.744748032	2.010357	3.598036
min	0.000000	0.000000	0.000000	0.000000	0.000000	2015-07-01 00:00:00	0.000000	0.000000
25%	0.000000	10.000000	0.000000	0.000000	72.250000	2016-03-21 00:00:00	2.000000	2.000000
50%	0.000000	45.000000	0.000000	0.000000	98.000000	2016-09-08 00:00:00	2.000000	3.000000
75%	1.000000	115.000000	0.000000	0.000000	132.000000	2017-03-02 00:00:00	2.000000	5.000000
max	1.000000	291.000000	13.000000	21.000000	230.000000	2017-09-14 00:00:00	12.000000	69.000000
std	0.445061	72.372885	0.304994	0.719954	46.218942	NaN	0.680556	2.745294

	is_repeated_guest	hotel	is_canceled
0	Not Repeated	Resort Hotel	0.229029
1	Not Repeated	City Hotel	0.311461
2	Repeated	Resort Hotel	0.037760
3	Repeated	City Hotel	0.105911

Lab One - Exploring Table Data¶

1. Business Understanding¶

Measure of Success¶

Real-World Examples and Benchmarks¶

2. Data Understanding¶

2.1 Loading the Dataset & Defining Data Types¶

Attributes Kept¶

Attributes Collected in Data¶

Categorical:¶

Numerical:¶

Attributes Dropped¶

Irrelvant:¶

Reundancy:¶

Low Impact:¶

Data Leakage:¶

2.2 Verifying Data Quality¶

Handling Missing and Duplicated Data¶

Missing Data:¶

Duplicated Data:¶

Converting to Appropriate Data Types¶

One Last Thing: Checking For Outliers¶

Data Understanding Summary¶

3. Data Visualization¶

3.1 Visualizing Basic Feature Distributions¶

Analysis:¶

Analysis:¶

Analysis:¶

Distribution Insights:¶

Analysis:¶

Analysis¶

Question 1: What is the impact of repeat guest status on cancellation rates across different hotel types?¶

Non-Repeated Guests Cancellation Rate¶

Repeated Guests Cancellation Rate¶

City Hotels Experience Higher Cancellation Rates¶

Question 2: How do cancellation rates change over time, and are there noticeable trends or patterns in cancellations for each type of hotel?¶

Analysis of Monthly Cancellation Rates¶

Trends Over Time:¶

Question 3: Which customer type is more likely to cancel?¶

Analysis¶

Cancellation Distributions by Customer Type¶

4. Additional Analysis¶

UMAP Dimensionality Reduction Methods¶